A Novel Intelligence Analysis Method Based on Graph Convolutional Neural Networks for Online Security of Power Distribution Networks

As the penetration of new energy power generation rises, various new energy systems increasingly impact the operation and management of distribution networks. Consequently, traditional methods for security analysis in distribution network systems are becoming inadequate for evaluating modern distribution networks. Based on this, this paper adopts a novel online safety intelligence analysis method for the distribution grid based on a graph convolutional neural network. This method takes the node voltage, current, power, and the location of distributed power generation as the node feature matrix, the connection relationship between nodes as the adjacency matrix, and the operation state of the distribution grid as the data label to establish a distribution grid operation graph database. It establishes a feature mining model of distribution grid operation data based on a graph convolutional neural network to complete the online intelligence analysis of the new-type distribution grid. It is proved that this method can realize faster, more accurate and more flexible online security intelligence analysis of the new distribution network.